theorycrafting: what's the currently optimal melee attack sequence?


General Discussion


afaict, the optimal melee attack sequence in PF2 is much as it was in PF1: rage, full attack with a falchion, and use Power Attack. Of course, in PF2-ese, "Power Attack" means "Vital Strike" and "be a giant totem barbarian" means "Power Attack."

That is, here's the best I can come up with. You (a) play a fighter, (b) take Barbarian Dedication (giant totem), Rager, Totem, Certain Strike, and Desperate Finisher, (c) wield a weapon with the forceful trait, and (d) be quick (via e.g. haste, Weapon Supremacy, or a speed weapon).

Then you do the following:

--) round 1: Rage -> Strike (extra action from quick) -> Certain Strike -> Certain Strike -> Certain Strike (reaction from Desperate Finisher)
--) rounds 2-3: Strike (extra action from quick) -> Certain Strike -> Certain Strike -> Certain Strike -> Certain Strike (reaction from Desperate Finisher)

The point is to maximize your flat damage bonuses, and hence maximize the effects of Certain Strike. It's super effective but super boring. The dice barely matter.

Overall DPR: 183.95 vs. AC 44 at 20th level (assuming Str 24, +5 falchion, crit on 19 or 20, +2d6 damage from property runes). An equivalent vanilla fighter gets 159.35 DPR with this tactic vs. the same AC, given that they're not sluggish and not spending an action to rage. So the fighter/barbarian breaks even in round 1 and pulls ahead afterwards.


3 people marked this as a favorite.

Actually, we have been explained why the old Power Attack wouldn't work in this version (I think it's in the old Fighter preview thread).
Then someone made the math showing that the new Power Attack (= Vital Strike) is not that good, either.


edit: this raises interesting issues that deserve its own post. In this one, I'm more interested in DPR theorycrafting.


1 person marked this as a favorite.
Pathfinder Starfinder Society Subscriber

My theorycrafting efforts around the release of the playtest also concluded that long Certain Strike + Forceful chains were the best DPR as well. I think what you suggest is the OP is complete and correct.

The only thing I have to add is that the rest of the game's math is fairly tight, leaving this "CS+F" strat as far and away the highest damage. Most martial classes cap out around 60-70 DPR, increased a little more if you include a barbarian dip (something I didn't have access to when I did my initial analysis). Monk has a combination that gets to a little over 100 DPR.

What that says to me is that Paizo will likely trim this strategy back for the final release.


Something I found that does pretty good at damage is going cleric negative energy gorum with sorcerer dedication to get magical striker
and then doing true strike into channel smite. True strike helps you hit a lot and channel smite gets the most damage for a hit.

Similar strat with fighter and power attack using a pick, with it's weapon crit effect having lots of dice means lots of damage on a crit and true strike helps you roll high for those crits.

I'm not sure what their DPR would be since I've never been good with true strike in the equation, but I believe these are some of the top DPR.

Paizo Employee Customer Service Representative

1 person marked this as a favorite.

Removed a post.

Avoid the use of profanity or other charged language.


4 people marked this as a favorite.
Pathfinder Adventure Path Subscriber
Chess Pwn wrote:


I'm not sure what their DPR would be since I've never been good with true strike in the equation, but I believe these are some of the top DPR.

Odds of each roll under true strike:

1 = 1 in 400 (0.25%)
2 = 3 in 400 (0.75%)
3 = 5 in 400 (1.25%)
4 = 7 in 400 (1.75%)
5 = 9 in 400 (2.25%)
6 = 11 in 400 (2.75%)
7 = 13 in 400 (3.25%)
8 = 15 in 400 (3.75%)
9 = 17 in 400 (4.25%)
10 = 19 in 400 (4.75%)
11 = 21 in 400 (5.25%)
12 = 23 in 400 (5.75%)
13 = 25 in 400 (6.25%)
14 = 27 in 400 (6.75%)
15 = 29 in 400 (7.25%)
16 = 31 in 400 (7.75%)
17 = 33 in 400 (8.25%)
18 = 35 in 400 (8.75%)
19 = 37 in 400 (9.25%)
20 = 39 in 400 (9.75%)


2 people marked this as a favorite.

I'd have to see some _very_ detailed math regarding the OP before I believe that this is optimal. I'm pretty sure that Giant Totem barbarian is a complete trap, so it's a bit odd to see it in an optimal DPR build.

My simulation models show that -1 to hit is certainly not worth the damage increase of even Giant Totem. I haven't updated those in a while, though.


NemisCassander wrote:

I'd have to see some _very_ detailed math regarding the OP before I believe that this is optimal. I'm pretty sure that Giant Totem barbarian is a complete trap, so it's a bit odd to see it in an optimal DPR build.

My simulation models show that -1 to hit is certainly not worth the damage increase of even Giant Totem. I haven't updated those in a while, though.

The deal is the reliance on certain strike to guarantee damage. Having a raging bonus of +6 damage per hit is really nice when you apply it 3 or 4 times a round.


Chess Pwn: Ah, thank you. That does make it pretty obvious.

Kinda shows the convolutions required in the system, though...


1 person marked this as a favorite.

Fighter with rogue dedication going Dread Striker -> double slice.

Max bluff, hope for the best. Having a 15-20 crit chance is the best you'll get on an appropriate challenge.


NemisCassander wrote:

I'd have to see some _very_ detailed math regarding the OP before I believe that this is optimal. I'm pretty sure that Giant Totem barbarian is a complete trap, so it's a bit odd to see it in an optimal DPR build.

My simulation models show that -1 to hit is certainly not worth the damage increase of even Giant Totem. I haven't updated those in a while, though.

I haven't run the math myself, but as a general point I'd bet you're right. But to paraphrase George Lucas, Certain Strike is the key to all of this.

Though your question was already answered, here's the detailed math. (Actually I did oversell it a bit: the OP numbers mistakenly ran the calculations against AC 43, not 44.)

Start with the assumptions in the OP: optimized 20th level fighter attacking AC 44. Your falchion averages 61 expected damage on a hit (6d10 + 2d6 + 7 str + 14 conditional), increasing to 67 on the second attack and 73 afterwards. On a failure, you do 35 damage from Certain Strike (6 + 2 + 7 + 14 + 6) on your second attack and 41 afterward.

So, here's the routine ([Q] abbreviates the extra action from quick):

Round One
[A] Rage
[Q] Strike at -0: 0.65 expected hits (0.55 hit, 0.1 crit) for 39.65 expected damage.
[A] Certain Strike at -5: 0.4 expected hits (0.3 hit, 0.4 crit) + 0.5 expected failures for a total of 44.3 expected damage (26.8 + 17.5).
[A] Certain Strike at -10: 0.1 expected hits (0.05 hit, 0.05 crit) + 0.5 expected failures for a total of 27.8 expected damage (7.3 + 20.5).
[R] Certain Strike at -10: 0.1 expected hits (0.05 hit, 0.05 crit) + 0.5 expected failures for a total of 27.8 expected damage (7.3 + 20.5).

total: 139.55 expected damage

Rounds Two & Three
[Q] Strike at -0: 0.65 expected hits (0.55 hit, 0.1 crit) for 39.65 expected damage.
[A] Certain Strike at -5: 0.4 expected hits (0.3 hit, 0.4 crit) + 0.5 expected failures for a total of 44.3 expected damage (26.8 + 17.5).
[A] Certain Strike at -10: 0.1 expected hits (0.05 hit, 0.05 crit) + 0.5 expected failures for a total of 27.8 expected damage (7.3 + 20.5).
[A] Certain Strike at -10: 0.1 expected hits (0.05 hit, 0.05 crit) + 0.5 expected failures for a total of 27.8 expected damage (7.3 + 20.5).
[R] Certain Strike at -10: 0.1 expected hits (0.05 hit, 0.05 crit) + 0.5 expected failures for a total of 27.8 expected damage (7.3 + 20.5).

total: 167.35 expected damage

Over the three rounds, the mean DPR is thus 158.08. By comparison, a vanilla fighter has a DPR of 143.15 when attacking this way; a fighter/barbarian without giant totem has a mean DPR of 151.75. In fact this scenario is actually abnormally bad for the giant totem barbarian, since sluggish 1 means they'll only hit on a natural 20 on their MAP - 10 attacks, negating the benefit of an expanded threat range. With a higher or lower baseline chance to hit, the giant totem's marginal advantage will be greater.


master_marshmallow wrote:

Fighter with rogue dedication going Dread Striker -> double slice.

Max bluff, hope for the best. Having a 15-20 crit chance is the best you'll get on an appropriate challenge.

I'm curious. Run some numbers?


okay so if we are going with lv20

Cleric has 20+6+5+1 = +32 to hit, enemy AC = 44
so cleric needs a 12 to hit with true strike that's apparently a 69.75% hit rate.
greatsword for 6d12+6+10d8+6 ~ 39+6+45+6 ~ 96*69.75% ~ 67 not factoring in crits.

Hmm... If this is correct it's a fair bit lower than expected. Though Part of the plan was to benefit from having a more magical weapon than normally possible with magical striker, but since +5 is what we have we lose out on that.

Fighter with a pick is really hoping for crits.
A normal power attack hit with a greatpick is 8d10 ~ 44 but on a crit that jumps to 17d12+17 ~ 127.5 from just the weapon. so true strike to up our crits should let us get this massive damage more often.

again, this "does better DPR" when you're not lv20 because then you benefit from being able to activate magical striker for accuracy and damage.


Ludovicus wrote:
Your falchion averages 61 expected damage on a hit (6d10 + 2d6 + 7 str + 14 conditional),

How are you getting 14 conditional damage? the rage is only giving you 8 conditional. And the non-giant would only have a +4 conditional.


Chess Pwn wrote:
Ludovicus wrote:
Your falchion averages 61 expected damage on a hit (6d10 + 2d6 + 7 str + 14 conditional),
How are you getting 14 conditional damage? the rage is only giving you 8 conditional. And the non-giant would only have a +4 conditional.

Ack, because I don't read carefully enough! I totally missed this:

Barbarian Dedication wrote:
For the purposes of the temporary Hit Points and damage bonus from Rage, your barbarian level is equal to half your level.

This lowers the non-giant-totem fighter/barbarian's mean DPR to 149.72. NemisCassander was right; giant totem remains a trap option. (Its mean DPR is 139.18, below the vanilla fighter.)

This also means that barbarian dedication is not worth it in practice, unless you are taking it for the feats. There are other things you can do with the feats, and the combined AC penalties from rage itself and the inability to take advantage of the fighter's master heavy armor proficiency are not justified by the marginal DPR increase.


There are several nitpicks I have with the calculations, but I think the idea is sound.

I/ Certain Strike percentages:

The expected failure rate for the 2nd hit and beyond should be 45%, not 50%.

At 40% accuracy, you hit on a 13+
You miss on 4 to 12 (9 numbers, 45% rate)
You crit miss on 1 to 3 (3 numbers, 15% rate)

At 10% accuracy, you hit on a 19 or 20
You miss on 10 to 18 aka 45%
You critically miss on 1 to 9 aka 45%.

Therefore the expected damage from Certain Strike should be slightly less (5% of a minimum hit with the Falchion) on the 2nd attack and beyond. As a general rule of thumb, if the accuracy of the attack is 50% or under, the failure rate is 45%, save for extreme cases where the enemy AC is astronomically high compared to your attack (like AC 21 vs. +0 to hit).

II/ Critical percentages:

On the first attack, if accuracy is 65%, crit chance should be 15%, not 10%, given PF2's critical math.

Speaking of Crits, the Fighter should have Savage Critical, making his minimum crit chance 10%, thus increasing his DPR.

III/ Accuracy:

A fully optimized Fighter with no buffs has +35 to hit (23 from proficiency, 7 from ability, 5 from item), but Titan Mauler makes you sluggish, bringing the bonus to +34.

Vs. AC 44, you hit on a 10, or 55% chance to hit. This reduction in accuracy affects the expected damage of your 1st and 2nd attack quite a bit.

IV/ Multiclassing disadvantages:

MCing Barbarian means you only get Rage damage as a level 10 Barb at most (+4 normally, +8 as Titan Mauler).

Does the +8 bonus to damage offset the attack bonus reduction?

V/ Putting it all together:

All in all, the Fighter MC Giant Totem Barb has a +34 to hit vs. AC 44, and on a hit he does 6d12 + 2d6 + 7 (Str) + 8 (rage) = 55 average damage.

Plug in the numbers and his attack sequence (Strike + 4 Certain Strike) yields 139.5 damage.

A normal Fighter does slightly less damage, but has +35 to hit. His average damage per hit is 47 damage, and his expected damage is 147.05 damage.

So it doesn't pay to MC Barb, since the accuracy loss is not offset by the +8 to damage. You also lose some action economy and some rounds off from fatigue when you rage.

Is there an alternative way to get conditional bonus to damage without going Barbarian?

Yes, there are several:

- Have your Bard cast Inspire Courage to get a +1, with Inspire Heroics, this could go up to +3. The bonus is small without adding in uncertainty into the equation.

- Multiclass Paladin and pick up Righteous Ally (Blade) and Blade of Justice. Only works on evil enemies, and requires an action each round unless the foe attacks your allies.

- Multiclass Cavalier and pick up Challenge. Most reliable way to get +6 conditional bonus to damage vs. a specific foe, but only lasts until the start of your next turn unless your allies don't help you against the foe.

Are any of these worth it? Under specific circumstances, Blade of Justice can last forever, meaning you don't lose actions to use it each turn, and it carries no penalties, so Blade of Justice will result in the best DPR if you're constantly fighting Evil enemies that ignore you and attack your allies each turn.

Challenge takes up an action each turn on average, and attacking with only 4 Strikes means that even with the added damage, you only deal 138.25 expected damage.

Best case is you just have the Bard cast Inspire Courage every turn and reap the benefits. In the absence of other buffs, the +1 to attack and damage spikes your expected damage up to 167.1, or 187.75 if the Bard succeeds on an Inspire Heroics check, or 209 if they critically succeed.

In comparison, a Fighter fully buffed with Heroism but no other conditional bonus to damage does 195.65 damage in a round with his 5 attacks.


Pramxnim wrote:
There are several nitpicks I have with the calculations, but I think the idea is sound.

I'm somehow getting slightly lower numbers: the vanilla fighter averages 138.05 DPR unassisted against AC 44. I think my math is consistent with everything you point out about certain strike and crit percentages; hmm.


Okay, the OP was wrong.

Or at least, it was wrong if unarmed strikes count as wielded weapons.* Wolf Jaw is a d8 agile attack that gets forceful when you're flanking the target, which makes multiple attacks with it extremely powerful at relatively high accuracy (which the fighter provides). Thus, take an optimized fighter with Monk Dedication, Basic Kata (wolf stance), Desperate Finisher, Monk's Flurry, Stance Savant, and Two-Weapon Flurry. (If you really wanted to, you could eke out a bit more damage by taking Rogue Dedication and Sneak Attacker, but I haven't included this in the numbers.) This enables:

Flurry of Blows -> Two-Weapon Flurry -> Two-Weapon Flurry -> Strike -> Two-Weapon Flurry

Unassisted, this sequence averages 135.3 DPR. But with flanking, the number jumps to 213.5; buffed with a critically successful Inspire Heroics, it becomes 308.4. A pit fiend has 300 hit points.

I'm now very curious if there's a way to top this.

____
* I take the rules to be ambiguous on this point. For the record, however, I should note that were I a GM, I would rule that unarmed strikes definitely do not count as wielded weapons for purposes of Two-Weapon Flurry, since the above build uses Two-Weapon Flurry essentially as a loophole to get around Flurry of Blow's once-per-round limitation. Presumably, that limitation was intended precisely to preclude shenanigans like this.


Ludovicus wrote:
For the record, however, I should note that were I a GM, I would rule that unarmed strikes definitely do not count as wielded weapons for purposes of Two-Weapon Flurry, since the above build uses Two-Weapon Flurry essentially as a loophole to get around Flurry of Blow's once-per-round limitation. Presumably, that limitation was intended precisely to preclude shenanigans like this.

Finally, note that even without Two-Weapon Flurry, Wolf Jaw still appears to be the best melee attack available if you're flanking.

As far as I can tell, the next-best alternative is a build that takes Certain Strike and Agile Grace in place of Two-Weapon Flurry, for:

Flurry of Blows -> Certain Strike -> Certain Strike -> Strike -> Certain Strike

For this, I'm getting 205.85 damage if you're flanking, and 297.4 for a critically successful inspire heroics.

Upshot: damagewise, the fighter/monk is the build to beat in PF2.


Chess Pwn wrote:


Hmm... If this is correct it's a fair bit lower than expected. Though Part of the plan was to benefit from having a more magical weapon than normally possible with magical striker, but since +5 is what we have we lose out on that.

Is there a reason to think that Magical Striker can't push a weapon to be an effective +6?


Are you using our formulas?

On mobile rn, but I was in need of a math thread.


Xenocrat wrote:
Chess Pwn wrote:


Hmm... If this is correct it's a fair bit lower than expected. Though Part of the plan was to benefit from having a more magical weapon than normally possible with magical striker, but since +5 is what we have we lose out on that.
Is there a reason to think that Magical Striker can't push a weapon to be an effective +6?

By RAW, none that I can see.


Ludovicus wrote:
Xenocrat wrote:
Chess Pwn wrote:


Hmm... If this is correct it's a fair bit lower than expected. Though Part of the plan was to benefit from having a more magical weapon than normally possible with magical striker, but since +5 is what we have we lose out on that.
Is there a reason to think that Magical Striker can't push a weapon to be an effective +6?
By RAW, none that I can see.

Agreed, I was surprised to see an apparent belief otherwise.


Ludovicus wrote:

Unassisted, this sequence averages 135.3 DPR. But with flanking, the number jumps to 213.5; buffed with a critically successful Inspire Heroics, it becomes 308.4. A pit fiend has 300 hit points.

I'm now very curious if there's a way to top this.

update: There is.

Community / Forums / Archive / Pathfinder / Playtests & Prerelease Discussions / Pathfinder Playtest / Pathfinder Playtest General Discussion / theorycrafting: what's the currently optimal melee attack sequence? All Messageboards

Want to post a reply? Sign in.
Recent threads in Pathfinder Playtest General Discussion